Today I used a project: https://github.com/brycedrennan/imaginAIry

This project is called ImaginAIry, which is a Python-based image generation project that uses AI technology to create and edit images. It can run an API server and the StableStudio web interface, supporting the generation of images through various control methods such as ControlNet, Canny Edge, and HED Boundary.

I wrote a sentence

prompts = [
    ImaginePrompt("fruit salad", tile_mode=True),
]
for result in imagine(prompts):
    display(result.img)

I asked it to change the makeup for the pearl girl and create a Halloween version.

p = ImaginePrompt(
    f"make it halloween",
    prompt_strength=15,
    init_image=LazyLoadingImage(url="https://raw.githubusercontent.com/brycedrennan/imaginAIry/master/assets/mask_examples/pearl000.jpg"),
    init_image_strength=1,
    model="edit"
)
prompts = [p]
for result in imagine(prompts):
    display(result.img)

It can also be used in combination with ControlNet:

imagine --control-image assets/indiana.jpg  --control-mode openpose --caption-text openpose "photo of a polar bear"

Compared to the Stable Diffusion WebUI, it is more suitable for me and other friends who don't like operating GUI programs.